In the search for promising drug candidates that will make their way to approval, drug developers sift through thousands of molecules to find the best one for clinical trials. To narrow it down, part of the process involves tracking where the drugs go in the body and how long they stay there. For small molecule drugs, there are ways to test multiple drug candidates within the same mouse. But scientists must test therapeutic antibody drugs exhaustively one at a time in mice. That’s because the structure of antibodies is so similar that it makes it nearly impossible to tell them apart.
Johannes vom Berg (pictured here) and his lab teamed up with Markus Seeger’s lab to blend complementary expertise to study flycodes in vivo.
Credit: Polina Mishchenko
However, new research led by Johannes vom Berg, a neuroimmunologist at the University of Zurich, and Markus Seeger, a microbiologist also at the University of Zurich, solved this problem by finding a way to identify individual antibodies within a mixture (1). The foundation of their advance is based on the use of unique short peptide tags attached to each antibody — called flycodes — that Seeger’s group previously designed (2). These novel flycodes finally allow researchers to recognize each antibody when they test them together in the same mouse. Their work could lead to fewer animals, lower costs, and faster timelines.
The team invented flycodes because traditional methods like using the trypsin enzyme to break up proteins at specific sequences have not been very successful for antibodies. This method would most likely produce peptides that are identical to each other with only a few unique peptides, making it difficult to identify the antibody. Therefore, Seeger’s lab designed flycodes to maximize sequence diversity and detectability.
As a first step, the team tested the tags on three clinical-stage antibodies used to treat autoimmune diseases and cancer. By comparing these antibodies to their untagged versions, they found that the flycodes didn’t change the antibodies’ pharmacokinetic properties. The team also confirmed that they could determine pharmacokinetic properties of individual antibodies within a larger pool of 23 clinical-stage antibodies, including immunoglobulin subclasses that target different diseases and locations in the body.
Beyond antibodies, the authors tested their tagging strategy with other proteins such as sybodies, which are synthetic single-domain antibodies designed to bind proteins. They found that they could detect the flycoded sybodies in mice and determine the molecules’ half-lives, hinting at the utility of flycodes with different proteins. “That’s the advantage of this approach. It’s not only limited to antibodies,” said Zhaohui Sunny Zhou, a protein chemist at Northeastern University who was not involved in this study.
The researchers also wanted to make sure that the flycodes could help them simultaneously track multiple antibody drug candidates to see if they reach their targets. To demonstrate proof of concept for these types of studies, they implanted mice with lung tumors that were engineered to overexpress human epidermal growth factor receptors (EGFRs), proteins frequently elevated in a number of cancer cells. They pooled together a sybody that bound EGFR, an antibody that bound EGFR, and antibodies that did not. When they introduced the mixture into the mouse, they were hoping to see that EGFR-binders ended up in the tumors and non-EGFR antibodies did not. Using the flycodes to detect the EGFR-binders, this was exactly what they found. “We were positively surprised that it worked so well,” said vom Berg.
However, adding tags to proteins has the possibility to affect protein structure and function. “Even if you change one or two amino acids, you can significantly change the properties,” said Zhou. To address this possibility, the researchers created 30 to 40 versions of the same antibody, each with a unique flycode. This could help mitigate any structural or functional changes due to a particular flycode, as any similarities in the same antibody’s behavior even though they have different tags could likely be attributed to the antibody itself.
We were positively surprised that it worked so well.
- Johannes vom Berg, University of Zurich
In total, these experiments used just 18 mice. “The potential to save animals is very big,” said vom Berg. Depending on the number of antibody candidates tested at once, this method reduces animal use by up to 100-fold.
Reducing animal use is a great endeavor in and of itself, but it also comes with other advantages: faster experiments and fewer variables. “This is a way to test a large number of biomolecules side by side in the most controlled setting that you can come up with, which is in the same mouse,” said vom Berg.
Vom Berg added that it’s important that this technology comes into play at the right time. “In the development path, there is a moment when this technique can be very useful,” he said. For example, Seeger said that this method could help rank a large number of antibody drug candidates to prioritize which ones move further down the pipeline. Zhou added, “I think it’ll be very useful for early-stage discovery.”
References
- Walter, J.D. et al. Flycodes enable simultaneous preclinical analysis for dozens of antibodies in single cassette–dosed mice. Proc Natl Acad Sci USA 122, e2426481122 (2025).
- Egloff, P. et al. Engineered peptide barcodes for in-depth analyses of binding protein libraries. Nat Methods 16, 421-428 (2019).
